Add files using upload-large-folder tool

README.md

@@ -0,0 +1,89 @@
+---
+library_name: motius
+pipeline_tag: other
+tags:
+- motion-to-text
+- motion-captioning
+- humanml3d
+- tm2t
+---
+
+<h1 align="center">TM2T Model Card</h1>
+
+<p align="center">
+  <strong>Tokenized reciprocal motion-language translation, reproduced as a standalone Motius M2T pipeline.</strong>
+</p>
+
+<p align="center">
+  <a href="https://arxiv.org/abs/2207.01696">Paper</a> |
+  <a href="https://ericguo5513.github.io/TM2T/">Project Page</a> |
+  <a href="https://github.com/EricGuo5513/TM2T">Original GitHub</a> |
+  <a href="https://huggingface.co/ZeyuLing/Motius-TM2T-HumanML3D">Motius Checkpoint</a>
+</p>
+
+TM2T is the ECCV 2022 reciprocal text-to-motion and motion-to-text method. The
+Motius release contains the HumanML3D VQ tokenizer, motion-to-text Transformer,
+vocabulary, statistics, and inference runtime. It does not import an original
+repository checkout.
+
+## Release Snapshot
+
+| Item | Value |
+| ---- | ----- |
+| Released task | M2T |
+| Motion representation | HumanML3D-263, 20 fps |
+| Motion tokenizer | 1,024-code VQ tokenizer |
+| Caption model | 4-layer encoder / 4-layer decoder Transformer |
+| Decoding | Beam search, beam size 2 |
+| Checkpoint | [`ZeyuLing/Motius-TM2T-HumanML3D`](https://huggingface.co/ZeyuLing/Motius-TM2T-HumanML3D) |
+| Pipeline | `motius.pipelines.tm2t.TM2TPipeline` |
+
+## Usage
+
+```python
+import numpy as np
+from motius.pipelines.tm2t import TM2TPipeline
+
+pipe = TM2TPipeline.from_pretrained(
+    "ZeyuLing/Motius-TM2T-HumanML3D",
+    bundle_kwargs={"device": "cuda"},
+)
+motion = np.load("sample.npy")  # denormalized HumanML3D-263
+caption = pipe.infer_m2t([motion], lengths=[len(motion)])[0]
+```
+
+## M2T Evaluation
+
+Full 4,402-clip evaluation follows the shared [HumanML3D M2T protocol](../tasks/m2t.md).
+Results are published only after the complete prediction set and metric artifact
+pass the population and sample-ID checks.
+
+| Samples | BLEU-4 | ROUGE-L | CIDEr | BERT F1 | R@1 | R@2 | R@3 | Matching |
+| ------: | -----: | ------: | ----: | -------: | --: | --: | --: | -------: |
+| 4,402 | - | - | - | - | - | - | - | - |
+
+## Motion Representation
+
+TM2T normalizes HumanML3D-263 features with its released training statistics.
+The VQ encoder removes four contact dimensions, maps each clip to discrete
+motion tokens, and the reciprocal Transformer translates those tokens to text.
+
+## Motius Components
+
+| Component | Path |
+| --------- | ---- |
+| Pipeline | `motius/pipelines/tm2t/pipeline.py` |
+| Bundle | `motius/models/tm2t/bundle.py` |
+| Runtime | `motius/models/tm2t/network.py` |
+| License | `motius/models/tm2t/LICENSE` |
+
+## Citation
+
+```bibtex
+@inproceedings{guo2022tm2t,
+  title={TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts},
+  author={Guo, Chuan and Zuo, Xinxin and Wang, Sen and Cheng, Li},
+  booktitle={European Conference on Computer Vision},
+  year={2022}
+}
+```

glove/our_vab_data.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73ed9b73ee5bc5ab683661c132b9f98fe305672dffdf70ceeac1bc447543d668
+size 10077728

glove/our_vab_idx.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d85dce83d2c27a92bb94bb51dc1f55a04fcaa328c359d23eccf648e296c16493
+size 79811

glove/our_vab_words.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fac88da1be6f00d36b72b88c9745c8ac35a94a7bd6ccf1fa3bab380faf8c2e0d
+size 67470

model_index.json

@@ -0,0 +1,9 @@
+{
+  "_class_name": "TM2TPipeline",
+  "_library_name": "motius",
+  "model_type": "tm2t",
+  "format": "motius-tm2t-artifact-v1",
+  "bundle_class": "motius.models.tm2t.TM2TBundle",
+  "pipeline_class": "motius.pipelines.tm2t.TM2TPipeline",
+  "api": "motius.pipelines.tm2t.TM2TPipeline.from_pretrained"
+}

t2m/M2T_EL4_DL4_NH8_PS/meta/mean.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bdb5ba69a3a9e34d71990db15bc535ebc024c8d95ddb5574196f96058faa7d3
+size 2232

t2m/M2T_EL4_DL4_NH8_PS/meta/std.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a5f7d60301c9465972fc225f8ad0ee8f957e7720431189123eb6d15873a9557
+size 2232

t2m/M2T_EL4_DL4_NH8_PS/model/finest.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b338050c5792505beae582c869cca84c06019a430909f1af7a3a12abd6d41c2
+size 385708975

t2m/M2T_EL4_DL4_NH8_PS/opt.txt

@@ -0,0 +1,40 @@
+------------ Options -------------
+batch_size: 64
+checkpoints_dir: ./checkpoints
+codebook_size: 1024
+d_inner_hid: 2048
+d_k: 64
+d_model: 512
+d_v: 64
+dataset_name: t2m
+dim_vq_dec_hidden: 1024
+dim_vq_enc_hidden: 1024
+dim_vq_latent: 1024
+dropout: 0.1
+eval_every_e: 5
+gpu_id: 1
+is_continue: False
+is_train: True
+label_smoothing: False
+lambda_beta: 1
+lambda_m2t: 1.0
+log_every: 50
+lr: 0.0002
+m2t_v3: False
+max_epoch: 300
+max_text_len: 20
+n_dec_layers: 4
+n_down: 2
+n_enc_layers: 4
+n_head: 8
+n_resblk: 3
+name: M2T_EL4_DL4_NH8_PS
+proj_share_weight: True
+q_mode: cmt
+save_every_e: 10
+save_latest: 500
+t2m_v2: False
+text_aug: False
+tokenizer_name: VQVAEV3_CB1024_CMT_H1024_NRES3
+unit_length: 4
+-------------- End ----------------

t2m/VQVAEV3_CB1024_CMT_H1024_NRES3/meta/mean.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bdb5ba69a3a9e34d71990db15bc535ebc024c8d95ddb5574196f96058faa7d3
+size 2232

t2m/VQVAEV3_CB1024_CMT_H1024_NRES3/meta/std.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a5f7d60301c9465972fc225f8ad0ee8f957e7720431189123eb6d15873a9557
+size 2232

t2m/VQVAEV3_CB1024_CMT_H1024_NRES3/model/finest.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:839db9413fd44e48c0ae13f4b49df72aa500b16c180af9c097b4ce7782589354
+size 503264244